Evaluating the influence of land use, landscape properties, precipitation and fish on aquatic ecosystem functioning and biodiversity through large temporal and spatial scale assessments across lakes and reservoirs

(1)

1 Serra Caiada, RN

EVALUATING THE INFLUENCE OF LAND USE, LANDSCAPE PROPERTIES, PRECIPITATION AND FISH ON AQUATIC ECOSYSTEM FUNCTIONING AND

BIODIVERSITY THROUGH LARGE TEMPORAL AND SPATIAL SCALE ASSESSMENTS ACROSS LAKES AND RESERVOIRS

(2)

UNIVERSIDADE FEDERAL DO RIO GRANDE DO NORTE CENTRO DE BIOCIÊNCIAS

DEPARTAMENTO DE ECOLOGIA

PROGRAMA DE PÓS GRADUAÇÃO EM ECOLOGIA

REGINA LÚCIA GUIMARÃES NOBRE

NATAL/RN 2019

(3)

REGINA LÚCIA GUIMARÃES NOBRE

Tese de doutorado apresentada ao programa de Pós-Graduação em Ecologia da Universidade Federal do Rio Grande do Norte, como parte dos requisitos para a obtenção do título de Doutor(a) em Ecologia.

Orientador(a): Dra. Luciana da Silva Carneiro

Coorientador(a): Dr. Adriano Caliman Ferreira Da Silva

NATAL/RN 2019

(4)

REGINA LÚCIA GUIMARÃES NOBRE BANCA EXAMINADORA

____________________________________________

Profa. Dra. Luciana da Silva Carneiro – Presidente da banca e orientadora

____________________________________________

Profa. Dra. Juliana Deo Dias - Membro interno UFRN

____________________________________________

Profa. Dra.Vanessa Becker - Membro interno UFRN

____________________________________________

Prof. Dr. Rafael Dettogni Guariento – Membro externo UFMS

____________________________________________

Prof. Dr. Rosemberg Menezes – Membro externo – UFPB

NATAL/RN 2019

(5)

O presente trabalho foi realizado com apoio da Coordenação de

Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - Código

de Financiamento 001

(6)

This work was supported by the Coordination for the Improvement of

Higher Education Personnel - Brazil (CAPES) - Financing Code 001

(7)

Nobre, Regina Lúcia Guimarães.

Evaluating the influence of land use, landscape properties, precipitation and fish on aquatic ecosystem functioning and biodiversity through large temporal and spatial scale

assessments across lakes and reservoirs / Regina Lúcia Guimarães Nobre. - Natal, 2020.

131 f.: il.

Tese (Doutorado) - Universidade Federal do Rio Grande do Norte. Centro de Biociências. Programa de Pós-graduação em Ecologia.

Orientadora: Profa. Dra. Luciana da Silva Carneiro.

Coorientador: prof. Dr. Adriano Caliman Ferreira da Silva.

1. Ciclagem de nutrientes - Tese. 2. Subsídios alóctonos _ Tese. 3. Fontes não pontuais de poluição _ Tese. 4. Carcaças de peixes - Tese. 5. Eutrofização - Tese. 6. Biodiversidade - Tese. I. Carneiro, Luciana da Silva. II. Silva, Adriano Caliman

Ferreira da. III. Universidade Federal do Rio Grande do Norte. IV. Título.

RN/UF/BSE-CB CDU 574

Universidade Federal do Rio Grande do Norte - UFRN Sistema de Bibliotecas - SISBI

Catalogação de Publicação na Fonte. UFRN - Biblioteca Setorial Prof. Leopoldo Nelson - -Centro de Biociências - CB

(8)

Agradeço à Universidade Federal do Rio Grande do Norte, e ao Departamento de Ecologia, por me acolher já há mais de 11 anos na instituição. À Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) por me proporcionar uma bolsa de pós-graduação.

Em especial, agradeço a Professora Dra. Luciana Carneiro, que desde a graduação orienta minha vida acadêmica. Obrigada por me incentivar a continuar trilhando os caminhos da ciência mesmo nos momentos em que pensei mudar de rumo, e obrigada pelos conselhos de vida! Agradeço fortemente ao meu coorientador, Professor Dr. Adriano Caliman, cuja dedicação em me orientar e ajudar a melhorar meu trabalho foram fundamentais para o desenvolvimento desta tese.

Agradeço a todos os professores e pós-doutores, alguns que passaram e outros que permanecem no departamento de Ecologia, por todo o conhecimento adquirido durante minha formação como Ecóloga. Especificamente agradeço à Miriam Plaza e Rafael Guariento por todas as críticas e instruções dadas para melhoria do meu trabalho durante a etapa de qualificação da tese. Aos professores que aceitaram compor banca de defesa do meu doutorado: Rosemberg

Menezes, Juliana Deo Dias, Vanessa Becker e Rafael Guariento. E a tantos outros, que em

vários momentos contribuíram de alguma forma para o desenvolvimento deste trabalho. Obrigada Adriana Carvalho, Ali Ger, Eduardo Venticinque.

À Camila Cabral, mulher, cientista, forte e inspiradora, que com garra e determinação enfrentou à coleta de dados em 100 lagos, do litoral ao semi-árido, e tornou essa pesquisa e tantas outras possíveis. Obrigada! Agradeço também a Letícia Quesado e demais envolvidos na coleta e processamento de dados relacionados ao projeto dos 100 lagos.

(9)

Meu sincero obrigada à Fulbright Brasil por me proporcionar um ano acadêmico no exterior repleto de vivências únicas. Agradeço ao professor Michael Vanni por me receber em seu laboratório na Miami University e por me possibilitar novas experiências acadêmicas, enriquecendo grandemente o meu período como doutoranda. Igualmente, agradeço à toda equipe do laboratório Vanni-Gonzaléz (Maria Gonzaléz, Amber Rock, Tanner Williamson,

Patrick Kelly), por todas as discussões semanais e conhecimentos trocados durante os lab meetings.

Obrigada aos meus amigos Paola Nobre, Natália Ross, Gabriel Brasil, Camilinha

Ataíde, Daniela Pereira, Diogo Átila, Natália Mabel, Janaíza Monte e Eldenir Vasconcelos

que sempre acreditaram em mim, independente das minhas escolhas.

Agradeço aos meus pais, Antônio Alfredo e Miriam Lúcia, fontes inesgotáveis de apoio e suporte, e à minhas irmãs Camila Cristina e Cíntia Mirela que mesmo longe, sempre estão aqui comigo compartilhando as dores e delícias da vida acadêmica.

Agradeço à Joris Guérin, meu “fellow fulbrighter”, colaborador e companheiro de vida. O seu incentivo foi definitivo para a conclusão dessa tese, Merci! Por fim, agradeço à sementinha que cresce dentro de mim, que me refez acreditar que vale a pena lutar por um mundo ecologicamente justo para as presentes e futuras gerações.

(10)

SUMMARY LIST OF FIGURES ...11 LIST OF TABLES ...14 ABSTRACT...15 RESUMO...16 GENERAL INTRODUCTION...17 Chapter 1 1. PRECIPITATION, LANDSCAPE PROPERTIES AND LAND USE INTERACTIVELY AFFECT WATER QUALITY OF TROPICAL LAKES AND RESERVOIRS ...29 1.1 Introduction...33 1.2 Methods...36 1.2 Results...47 1.4 Discussion...54 1.5 Conclusion...56 1.5 References... 58 Chapter 2 2. PHYTOPLANKTON COMMUNITY COMPOSITION AND STRUCTURE DIFFERS AMONG TROPICAL LAKES AND RESERVOIRS...68

2.1 Introduction...70 2.2 Methods...74 2.2 Results...80 2.4 Discussion...86 2.5 References...90 Chapter 3

(11)

3. FISH, INCLUDING THEIR CARCASSES, ARE NET NUTRIENT SOURCES TO

THE WATER COLUMN OF AN EUTROPHIC LAKE...99

3.1 Introduction...101 3.2 Methods...101 3.3 Results...103 3.4 Discussion...104 3.5 References...107 FINAL CONSIDERATIONS ...109 APPENDIX A ...112 APPENDIX B ...118

(12)

11

LIST OF FIGURES Chapter 1

Figure 1 – Panel a show the studied lakes sampled on Rio Grande do Norte, Brazil. Natural lakes

are represented by white dots and reservoirs by black dots. Land cover types were grouped according to the purposes of this study. Natural vegetated areas refers to natual forest, savanna and grassland formations. Anthropogenic land use encompasses farming (agriculture/pasture) and urban land uses. Original data was available through Project MapBiomas - Collection 3 of Brazilian Land Cover & Use Map Series. Map was build using ArcGIS 10.5.1 (ESRI, 2017). Panel

b is presenting the average monthly precipitation from 2010 to 2012 in the studied lakes calculated

from interpolation of the seven INMET automatic weather stations (stars). The dashed line depicts the semi-arid delimitation according to SUDENE, 2017. To the right of the line is the humid region and to the left is the semi-arid.

Figure 2 – PCA biplot of water quality parameters (TP, TN and Chl-a) from the 98 lakes studied

at Rio Grande do Norte, Brazil.

Figure 3 – Regression tree showing the relationships between water quality impairment (WQI)

and precipitation, landscape and land use properties for the studied lakes. WQI is a composite response variable originated from the PCA using TN, TP and Chl-a. High values indicates low water quality (higher concentration of nutrients and Chl-a) and low values indicate less impaired water quality. Each split in the tree represents a Yes or No answer to the condition stated in each node box. Circles represents terminal nodes and it lists the mean value of the WQI and the number of lakes belonging to this group. AntBA: % of anthropogenic land use on the buffer extent; AntCA: % of anthropogenic land use on the catchment area extent, CVprecip: coefficient of variation of precipitation; BA:Vol: buffer area to volume ratio.

Chapter 2

Fig. 1 – Map indicating the location of the 98 lakes (white dots) and reservoirs (black dots)

distributed across the state of Rio Grande do Norte, Brazil. The dashed line depicts the semi-arid delimitation according to SUDENE (2017). To the right of the line is the coastal humid region and to the left is the semi-arid.

(13)

12

Fig. 2 Comparisons of landscape and land use properties among lakes (n=30) and reservoirs

(n=68). a. Extension of anthropogenic land use on the catchment scale; b. Extension of anthropogenic land use on the buffer scale; c. Catchment area to volume ratio; d. Buffer area to volume ratio; e. lake perimeter to volume ratio. Differences on the average values of the mentioned variables were tested with Student’s t test. Columns and vertical bars depict the mean and ± 95% confidence intervals.

Fig. 3 Comparisons of physio/chemical parameters among lakes (n=30) and reservoirs (n=68). a.

TN; b. TP; c. N:P ratio d. Depth of euphotic zone; e. pH; f. Temperature. Differences on the average values of TN, TP, N:P, depth of euphotic zone and temperature were tested with Student’s t test. Columns and vertical bars depict the mean and ± 95% confidence intervals. pH was tested through a nonparametric Mann–Whitney test. Boxes and horizontal bars represent the interquartile range, median and the data range, respectively.

Fig. 4 Comparisons among lakes (n=30) and reservoirs (n=68) regarding the local average

phytoplankton a. species diversity (Simpson’s index); b.species evenness; c. Species richness; as well as d. phytoplankton biovolume; e. cyanobacteria biovolume; f. relative cyanobacteria biovolume; g. toxic cyanobacteria biovolume and h. relative toxic cyanobacteria biovolume. Differences on the average values of a, c, f and g were tested with Mann – whitney test. Boxes and horizontal bars represent the interquartile range, median and the data range, respectively. Averages of all other variables were teste through student’s t test. Columns and vertical bars depict the mean and ± 95% confidence intervals.

Fig 5 –Phytoplankton gamma diversity (regional species richness) for lakes (n=30), reservoirs

(n=68) and all aquatic systems combined (n=98). Graphic shows individual-based rarefaction curves for the average (solid lines) accumulated phytoplankton species richness and their ± 95%confidence intervals (shaded areas). The scale of abundance is shown until that difference in species richness for lakes and reservoirs is stabilized

Fig. 6 - Nonmetric multidimensional scaling (NMDS) plots based on a Jaccard similarity matrix

(14)

13

Chapter 3

Fig. 1 - Conceptual model showing the fluxes of nutrients derived from fish (sources to the water

column, solid arrows) and stored in fish biomass (sinks from the water column, dashed arrows). Fish can act as nutrient sinks when they remove nutrients from circulation and store them in their bodies for growth, or when fish die and nutrients are stored in recalcitrant tissues. On the other hand, they act as sources of nutrients to the water column when they are releasing nutrients through excretion and remineralization of nutrients stored in carcasses. Piscivory is omitted here for simplicity.

Fig. 2 - Individual and interactive effects of fish size and temperature on patterns of mass

remaining for (A,B) nitrogen (C,D) phosphorus mass in decomposing fish carcasses throughout time. Values are expressed as the proportion of initial mass in carcasses. Data for young-the-year (YOY) and adult fish are shown in separate panels for clarity. The transition between the white and gray areas in the panels indicate the half-lives of mass decay. Data points are mean (n = 3) and error bars are ± 1SEM.

Fig. 3 - Individual and interactive effects of fish size (S) and temperature (T) on time integrated

fish carcass decay coefficients (k) for (A) nitrogen and (B) phosphorus. Data points are means (n = 3) and error bars are ± 1SE. Different letters above treatments indicate significant statistical differences among treatments. Bold p-values depict significant statistical effects (Tukey’s pos-hoc test; p < 0.05).

Fig. 4 - Twenty-year time series showing the dynamics of phosphorus (blue line) and nitrogen (red

line) stored in carcasses over time. The large peak occurring during winter is due to high mortality (high carcass production) rate of all age classes during this time, whereas the smaller peaks occurring in late July-August are due to production of carcasses from high mortality of young-of-year fish.

Fig. 5 - Annual fluxes of phosphorus (A) and nitrogen (B) for net biomass production (white

(15)

14

LIST OF TABLES General Introduction

Table 1 - Examples of internal and external processes related to nutrient cycling in freshwaters....22

Chapter 1

Table 1 – Overview of landscape and land use general properties (mean ± SD & range) for lakes studied. Anthropogenic land use was composed by the combination of agriculture/pasture and urban/developed land uses, which are complementary to the forested/vegetated land use. CA depicts catchment area...37 Table 2 – Correlation matrix among the main variables studied. Significant correlations (Pearson correlation, p≤0.05) are highlighted in bold. Non-significant correlations are in light grey.

Table 3 – Description of variables used in the RTA model.

Table 4 – Alternative RTA Models ran with pairwise combination of Totprecip & CVprecip and LP:Vol & BA:Vol to select the best fit model. The model that presented the combination of higher R2 _{and lowest RMSE was chosen as the best fit model (highlighted in bold).}

Chapter 2

Table 1 – Morphometric characteristics for sampled lakes and reservoirs...76

Chapter 3

Table 1 – Q10 values for N and P decay rates through 38 days of the laboratory experiment...104 Table 2 – Comparison of different fluxes from sediments to water in Acton Lake...107

(16)

15

ABSTRACT: Nutrient cycling is a fundamental ecosystem service as it provides an adequate

balance of elements that are necessary for life. In freshwaters, the balance of nitrogen (N) and phosphorus (P) are of special interest as they often limit or control primary production and biomass formation. While the availability of these nutrients is fundamental for the maintenance of biodiversity and productivity of freshwaters, their excess can lead to eutrophic conditions that are associated with impaired water quality and biodiversity loss. The nutrient balance in freshwaters can potentially be affected by a variety of biotic and abiotic, external and internal pathways. In this thesis, two frameworks were explored. First, a spatial framework focused on external processes, where we investigated the direct and indirect effects that land use (i.e. type, extent), precipitation and landscape properties (i.e. lake origin, lake and catchment absolute and relative size and geomorphology) have on biotic and abiotic properties of freshwater systems. More specifically, in chapter one we evaluated, across 98 tropical lakes and reservoirs, the individual and interactive effects of land use, precipitation and landscape properties on patterns of water quality parameters (N, P and chlorophyll-a). In chapter two, we characterized the 98 lakes as natural or artificial and compared them regarding the landscape properties of their surroundings, their morphometry, and their physico/chemical characteristics to verify whether those factors can be associated with average patterns of phytoplankton community structure at both local and regional scales. The second approach, presented in Chapter 3, was a long-term temporal framework focused on internal processes related to nutrient cycling where we assessed whether an omnivorous fish with high biomass and growth rate is a source or sink of N and P to the pelagic zone of a temperate eutrophic lake, at various time scales ranging from days to 20 years.

Key-words: nutrient cycling, land use, allochthonous subsidies, nonpoint source pollution, fish

(17)

16

AVALIANDO OS EFEITOS DE PROPRIEDADES DO USO DO SOLO E DA PAISAGEM, DO CLIMA E DE PEIXES NO FUNCIONAMENTO E BIODIVERSIDADE

DE ECOSSISTEMAS AQUÁTICOS AO LONGO DE LARGAS ESCALAS ESPAÇO-TEMPORAIS EM LAGOS E RESERVATÓRIOS

RESUMO: A ciclagem de nutrientes é um serviço ecossistêmico fundamental, pois proporciona

um equilíbrio adequado dos elementos necessários à vida. Nos ecossistemas de água doce, o balanço de nitrogênio (N) e fósforo (P) é de especial interesse, pois frequentemente estes elementos limitam ou controlam a produção primária e a formação de biomassa. Embora a disponibilidade desses nutrientes seja fundamental para a manutenção da biodiversidade e produtividade dos ecossistemas limnéticos, seu excesso pode levar a condições de eutrofização que estão associadas à deterioração da qualidade da água e à perda de biodiversidade. O balanço de nutrientes nos sistemas de água doce pode ser potencialmente afetado por uma variedade de vias bióticas e abióticas, externas e internas. Nesta tese, duas abordagens foram exploradas, uma abordagem espacial, focada em processos externos, onde foram investigados os efeitos diretos e indiretos do uso da terra (i.e. tipo, extensão), precipitação e propriedades da paisagem (tamanho absoluto e relativo e geomorfologia do lago e bacia) nas propriedades bióticas e abióticas de ecossistemas aquáticos. Mais especificamente, no capítulo um, avaliamos em 98 lagos e reservatórios tropicais os efeitos individuais e interativos do uso da terra, precipitação e propriedades da paisagem sobre parâmetros relacionados à qualidade da água (N, P e clorofila-a). No capítulo dois, caracterizamos os 98 lagos como naturais ou artificiais e os comparamos em relação às propriedades da paisagem em seu entorno, sua morfometria e suas características físico-químicas para verificar se esses fatores podem ser associados a padrões médios da estrutura da comunidade fitoplanctônica em escalas locais e regionais. A segunda, é uma abordagem temporal de longo prazo, focada em processos internos relacionados ao ciclo de nutrientes, onde avaliamos se um peixe onívoro com alta biomassa e alta taxa de crescimento é uma fonte ou sumidouro de N e P para a zona pelágica de um lago temperado eutrófico em várias escalas de tempo variando de dias a anos.

Palavras – chave: Ciclagem de nutrientes, uso do solo, subsídios alóctones, fontes não pontuais

(18)

17

EVALUATING THE INFLUENCE OF LAND USE, LANDSCAPE PROPERTIES, CLIMATE AND FISH ON AQUATIC ECOSYSTEM FUNCTIONING AND BIODIVERSITY THROUGH LARGE TEMPORAL AND SPATIAL SCALE

ASSESSMENTS ACROSS LAKES AND RESERVOIRS

GENERAL INTRODUCTION

Nutrient cycling, i.e. the flux of nutrients between organisms, habitats or ecosystems is a fundamental ecosystem service as it provides an adequate balance of elements that are necessary for life (Vanni, 2002). In freshwaters, the research on nutrients balance is mainly focused on nitrogen (N) and phosphorus (P) as the supply rate of this key elements often limits or control primary production and biomass formation (Paerl, 2009). An adequate balance of nutrients is fundamental for the maintenance of productivity, resources availability and biodiversity of freshwaters, while the excess of nutrients can lead to eutrophic conditions that are associated with increased phytoplankton biomass, the abundance of nuisance and toxic algae, increased water turbidity, oxygen depletion and fish deaths, impairing freshwater quality and compromising its biodiversity and economic value (Carpenter et al., 1998; Catherine et al., 2010).

Ecosystems are connected by spatial flows of matter, energy and organisms across their physical boundaries ( Elton 1927, Polis & Hurd, 1995; Loreau, Mouquet and Holt 2003, Ballinger & Lake, 2006; Gratton, Donaldson & Zanden, 2008). A great proportion of ecosystems are thus subsidized via external processes. Freshwaters are especially subject to allochthonous subsidies because of gravity and its concave profiles in the landscape, which enables the flux of materials from the catchment areas into the water bodies (Vanni et al., 2005; Vannote et al., 1980). The role of internal processes on nutrient cycling via animals (e.g. fish) has also been well documented, once animals consume, store, release and translocate nutrients at multiple scales, and at ecologically relevant rates, having potential to affect nutrient cycling, water quality and ecosystem

(19)

18

productivity (Barton et al., 2013; Beasley et al., 2012; Subalusky and Post, 2018; Vanni, 2002; Vanni et al., 2006; Williamson et al., 2018). Thus, nutrient balance in freshwater systems can be affected by a variety of internal and external processes (Table 1).

In this thesis, two frameworks were explored. The first one is a spatial framework focused on external processes, where we investigated the individual and interactive effects of land use (i.e. type, proximity and extent) landscape properties (i.e. lake origin, lake and catchment absolute and relative size and geomorphology) and precipitation on water quality indicators and on local and regional taxonomic phytoplankton diversity, by using a dataset of 98 perennial natural lakes and reservoirs (hereafter in this section, lakes) located in Rio Grande do Norte, northeast Brazil. Their location expands from the coastal to the semi-arid region, with climate varying from humid to semi-arid and trophic conditions varying from oligotrophic to hypereutrophic. Furthermore, these 98 lakes are included in different contexts of land occupation and use, from pristine areas to urban concentrations, farming and livestock production areas, among others (Cabral et al., 2019; Junger et al., 2019).

The second approach is a temporal framework focused on internal processes related to nutrient cycling where we assess whether and how the long-term variation of population dynamics of an omnivorous fish with high biomass and growth rate act as a source or sink of N and P to the pelagic zone of a temperate eutrophic lake, at various time scales ranging from days to years. To accomplish this, we used a population-level model for gizzard shad (Dorosoma cepedianum) over a 20-year period in Acton Lake, a eutrophic reservoir in southwest Ohio, U.S.A (Williamson et al., 2018).

I.I External Process - The influence of catchment land use and its interactions with lake and landscape properties and precipitation

(20)

19

Earth’s surface has experienced large scale transitions from natural landscapes to altered land covers, through the creation and expansion of urban areas and the conversion of pristine lands to agriculture, pasture, and urban landscapes (Foley et al., 2005; Seto et al., 2012). Agriculture and urbanization are amongst the most intensive types of land use. As of 2015, farming lands alone, occupy a combined area of approximately 34% of the Earth’s ice free land surface (Carpenter et al., 2011; Ramankutty et al., 2008) overlapping the amount of forested areas (about 31%), while urban areas covered about 2.85% of land surface (The World Bank, 2019).

Over the past years, the impacts of human alterations on natural environments have become a rising research topic on environmental sciences because of the growing recognition of the adverse effects of land use on global ecosystems. Thus, anthropogenic land uses represent a potential threat to the provisioning and maintenance of ecosystem services and functions. Particularly, recent studies have shown an intimate relationship between the development of anthropogenic land use types and impaired water quality (Doubek et al., 2015; Jeppesen et al., 1999; Wang et al., 2008).

Agriculture and urban impervious surfaces are one of the primary sources of N and P to freshwaters (Paul and Meyer, 2008; Vanni et al., 2011). This artificial enrichment of N and P is the main threat to water quality worldwide (Carpenter et al., 2011) causing severe problems as toxic algal blooms, reduced oxygen on water column, fish death among others. These nutrients come from sewages, fertilizers, pesticides and detergents. Thus, land use can be a major factor controlling the export of nutrients from the catchment of inland aquatic systems, inducing changes in resources availability and nutrient ratios entering freshwaters, thus potentially affecting water quality and provoking changes in freshwater communities.

The effects of lands use on water quality and biodiversity can be better understood when combined with complementary environmental parameters such as climatic and morphometric

(21)

20

variables (Bucak et al., 2018; Catherine et al., 2010; Knoll et al., 2015). Landscape properties such as topography (i.e. declivity), the absolute and relative size of catchments and their lakes (Bremigan et al., 2008; Hayes et al., 2015; Vannote et al., 1980) and the origin (i.e. natural or artificial) of the aquatic systems (Dodson et al., 2006; Nielsen et al., 2012) are factors that can influence the transport rates of allochthonous material to waterbodies, as well as determine how sensible these systems are to external inputs of materials.

For instance, differences in lake morphometry can arise when the waterbodies are originated from different processes, such as natural lakes and reservoirs. While natural lakes are created from geologic time and forces (Thornton, 1984), reservoirs are newer ecosystems constructed in its majority in the last 60 years aiming to store water for multiple purposes (Soballe and Kimmel, 1987; Thornton, 1984). Compared to lakes, reservoirs usually present greater catchment size because dams are strategically built along higher order rivers to allow damming of water from a large catchment area. Reservoirs are also built in places with more rugged topography, and therefore, they usually also present higher depths (thus, greater volume), perimeter and consequently higher catchment area to lake volume (CA:LV) and perimeter to lake volume (LP:LV) ratios. Those features can lead to a higher potential influence from allochthonous materials entering the reservoirs (Doubek and Carey, 2017; Knoll et al., 2015; Menezes et al., 2018; Thornton, 1984), which in turn can affect lake productivity and alter patterns of local and regional plankton diversity (e.g. species richness, abundance and composition) (Cabral et al., 2019; Doubek et al., 2015; Mantzouki et al., 2015; Schindler, 2006).

Additionally, climatic characteristics such as the patterns of precipitation can also interact with landscape properties and land use to affect water quality and biodiversity. For instance, the magnitude and variability of precipitation can interfere by intensifying (e.g. via soil percolation or

(22)

21

surface runoff) or attenuating (e.g. via dilution) the land use effects on aquatic systems (Jeppesen et al., 2015). Thus, the interaction of these factors (land use, lake and landscape properties, precipitation) can reflect on the concentration of nutrients in the water which will affect its quality and the aquatic ecosystem productivity and biodiversity.

I.II Internal Process – the role of fish as source or sink of nutrients

Differently from allochthonous inputs of nutrients to water bodies, which are strictly sources of nutrients, animals, like fish can act as sources or sinks of nutrients at the ecosystem scale. A fish population can act as a source of nutrients when it is releasing nutrients in available forms to other members of the ecosystem, and as a sink when they are removing nutrients from the circulation in the ecosystem (Kitchell et al., 1975; Vanni et al., 2013). Most studies on nutrient cycling by fish (and other aquatic animals) have focused mostly on their role as a nutrient source, mainly through excretion (Atkinson et al., 2017; Subalusky and Post, 2018). However, because fish can represent a large proportion of animal biomass in many ecosystems, because they are long-lived compared to other organisms, and because their bodies contain recalcitrant tissues like bones and scales, it has been suggested that fish can act as a nutrient sink in pelagic freshwaters (Sereda et al., 2008).

The main ways in which a fish population can be a nutrient sink are 1) if its biomass increases, i.e., when growth and reproduction exceed mortality, 2) if emigration from the ecosystem exceeds immigration to that ecosystem, and 3) if nutrients stored in carcasses are not mineralized back to the water column, but rather remain stored in sediments in a recalcitrant form for a long time (Vanni et al., 2013). Therefore, the role of fish as a source or sink of nutrients may depend on the population dynamics, relative proportion between different life stages and conditions that control the decomposition of the carcasses (e.g. temperature).

(23)

22

Table 1 –Examples of internal and external processes related to nutrient cycling in freshwaters.

Process References

Internal Processes

*Animal – translocation, recycling (e.g via excretion, consumption, decomposition)

(Atkinson et al., 2017; Beasley et al., 2012; Schaus et al., 1997; Sereda et al., 2008; Subalusky et al., 2017; Vanni, 2002; Vanni et al., 2013; Williamson et al., 2018)

Re-suspension of sediments by wind and bioturbation

(Adamek and Maesalek, 2013; Boqiang et al., 2004; Forsberg, 1989; Kristensen et al., 1992; Matsuzaki et al., 2007; Sondergaard et al., 1992)

External Processes

*Nutrient runoff from watershed (Carpenter et al., 1998; Jeppesen et al., 1999; Ometo

et al., 2000; Vanni et al., 2011; Nielsen et al., 2012;

Knoll et al., 2015; Bucak et al., 2018)

Atmospheric deposition (Bergstrom et al., 2005; Elser et al., 2009; Jassby et al., 1994)

Transport of nutrients by migration (Mitchell and Lamberti, 2005; Naiman et al., 2002; Vanni, 2002)

(24)

23

Therefore, from the above, the main objectives of this thesis were; (1) to evaluate, through a large scale spatial assessment, the effects that different aspects of land use and landscape properties (e.g. lake and catchment morphometry) have on biotic and abiotic properties of tropical lakes and reservoirs and whether and how such effects can be interactively dependent among each other and with precipitation; and (2) to understand through a large-scale temporal approach, whether and how an omnivorous fish is a source or sink of nutrients through its long-term population dynamics, nutrient excretion and carcass decomposition in a temperate eutrophic reservoir. In order to do that, this thesis was structured in three chapters. The specific goals pertaining to each chapter are presented below.



Chapter 1 - Precipitation, landscape properties and land use interactively affect water

quality of tropical freshwaters

-

The aim of this study is to evaluate the effects of landscape properties (morphometric measurements of lakes and their catchments), precipitation patterns and land use properties (extent and proximity of the land use to water bodies) on water quality of 98 natural and artificial lakes in northeast Brazil.



Chapter 2 - Phytoplankton community composition and structure differs among

tropical lakes and reservoirs – The goal of this chapter is to characterize lakes and

reservoirs regarding the landscape properties of their surrounding areas (i.e. catchment size and catchment land use absolute and relative extent), their morphometry (i.e. ecosystem size, lake perimeter to lake volume ratio - LP:LV, and catchment area to lake volume ratio – CA:LV) and their physico/chemical characteristics, and verify whether those factors can be associated with average patterns of phytoplankton community structure at both local and regional scales across 98 lakes and reservoirs in Northeast Brazil.

(25)

24



Chapter 3 - Fish, including their carcasses, are net nutrient sources to the water

column of a eutrophic lake - In this paper, our goal is to assess whether a fish population

with high biomass and growth rate is a source or sink of nitrogen and phosphorus to the pelagic zone of a eutrophic lake, at various time scales ranging from days to 20 years.

REFERENCES

Adamek, Z., and Maesalek, B. (2013). Bioturbation of sediments by benthic macroinvertebrates and fish and its implication for pond ecosystems : a review. Aquacult Int, 1–17.

doi:10.1007/s10499-012-9527-3.

Atkinson, C. L., Capps, K. A., Rugenski, A. T., and Vanni, M. J. (2017). Consumer-driven nutrient dynamics in freshwater ecosystems : from individuals to ecosystems. Biol. Rev. 92, 2003–2023. doi:10.1111/brv.12318.

Ballinger, A., and Lake, P. S. (2006). Energy and nutrient fluxes from rivers and streams into.

Mar. Freshw. Res. 57, 15–28. Available at: http://www.publish.csiro.au/?paper=MF05154.

Barton, P. S., Cunningham, S. A., Lindenmayer, D. B., and Manning, A. D. (2013). The role of carrion in maintaining biodiversity and ecological processes in terrestrial ecosystems.

Oecologia, 761–772. doi:10.1007/s00442-012-2460-3.

Beasley, J. C., Olson, Z. H., and DeVault, T. L. (2012). Carrion cycling in food webs : comparisons among terrestrial and marine ecosystems. oikos 121, 1021–1026. doi:10.1111/j.1600-0706.2012.20353.x.

Bergstrom, A., Blomqvist, P., and Jansson, M. (2005). Effects of atmospheric nitrogen

deposition on nutrient limitation and phytoplankton biomass in unproductive Swedish lakes.

Limnol. Oceanogr. 50, 987–994.

Boqiang, Q. I. N., Weiping, H. U., and Guang, G. A. O. (2004). Dynamics of sediment resuspension and the conceptual schema of nutrient release in the large shallow Lake Taihu , China. Chinese Sci. Bull. 49, 54–64. doi:10.1360/03wd0174.

Bremigan, M. T., Soranno, P. A., Gonzalez, M. J., Bunnell, D. B., Arend, K. K., Renwick, W. H., et al. (2008). Hydrogeomorphic features mediate the effects of land use / cover on reservoir productivity and food webs. Limnol. Oceanogr. 53, 1420–1433.

doi:10.4319/lo.2008.53.4.1420.

Bucak, T., Trolle, D., Tav, Ü. N., Çak, A. İ., Özen, A., Jeppesen, E., et al. (2018). Science of the Total Environment Modeling the effects of climatic and land use changes on phytoplankton and water quality of the largest Turkish freshwater lake : Lake Bey ş ehir. Sci. Total

Environ. 621, 802–816. doi:10.1016/j.scitotenv.2017.11.258.

Cabral, C. R., Guariento, R. D., Ferreira, F. C., Amado, A. M., Nobre, R. L., Carneiro, L. S., et al. (2019). Are the patterns of zooplankton community structure different between lakes and

(26)

25

reservoirs? A local and regional assessment across tropical ecosystems. Aquat. Ecol. 53, 335–346.

Carpenter, S., Caraco, N. F., Correll, D. L., Howarth, R. W., Sharpley, A. N., and Smith, V. H. (1998). Nonpoint Pollution of Surface Waters with Phosphorus and Nitrogen. Issues Ecol. 4, 1–12. doi:1092-8987.

Carpenter, S. R., Stanley, E. H., and Vander Zanden, M. J. (2011). State of the World’s

Freshwater Ecosystems: Physical, Chemical, and Biological Changes. Annu. Rev. Environ.

Resour. 36, 75–99. doi:10.1146/annurev-environ-021810-094524.

Catherine, A., Mouillot, D., Escoffier, N., Bernard, C., and Troussellier, M. (2010). Cost effective prediction of the eutrophication status of lakes and reservoirs. freshwater, 2425– 2435. doi:10.1111/j.1365-2427.2010.02452.x.

Dodson, S. I., Everhart, W. R., Jandl, A. K., and Krauskopf, S. J. (2006). Effect of watershed land use and lake age on zooplankton species richness. Hydrobiologia 579, 393–399. doi:10.1007/s10750-006-0392-9.

Doubek, J. P., and Carey, C. C. (2017). Catchment , morphometric , and water quality

characteristics differ between reservoirs and naturally formed lakes on a latitudinal gradient in the conterminous United States. Inl. Waters 7, 171–180.

doi:10.1080/20442041.2017.1293317.

Doubek, J. P., Carey, C. C., and Cardinale, B. J. (2015). Anthropogenic land use is associated with N-fixing cyanobacterial dominance in lakes across the continental United States.

Aquat. Sci. 77, 681–694. doi:10.1007/s00027-015-0411-x.

Elser, J. J., Anderseb, T., Baron, J. S., Bergstrom, A., Jansson, M., Kyle, M., et al. (2009). Shifts in Lake N : P Stoichiometry and Nutrient Limitation Driven by Atmospheric Nitrogen Deposition Shifts in Lake N : P Stoichiometry and Nutrient Limitation Driven by Atmospheric Nitrogen Deposition. Science (80-. ). 326, 835–837.

doi:10.1126/science.1176199.

Foley, J. A., Defries, R., Asner, G. P., Barford, C., Bonan, G., Carpenter, S. R., et al. (2005). Global Consequences of Land Use. Science (80-. ). 309, 570–575.

Forsberg, C. (1989). Importance of sediments in understanding nutrient cyclings in lakes.

Hydrobiologia, 263–277.

Gratton, C., Donaldson, J., and Zanden, M. J. Vander (2008). Ecosystem Linkages Between Lakes and the Surrounding Terrestrial Landscape in Northeast Iceland. Ecosystems 11, 764– 774. doi:10.1007/s10021-008-9158-8.

Hayes, N. M., Vanni, M. J., Horgan, M. J., and Renwick, W. H. (2015). Climate and land use interactively affect lake phytoplankton nutrient limitation status. Ecology 96, 392–402. doi:10.1890/13-1840.1.

Jassby, A. D., Reuter, J. E., Axler, R. P., Goldman, C. R., and Hackley, S. H. (1994).

Atmospheric Deposition of Nitrogen and Phosphorus in the Annual Nutrient Load of Lake Tahoe ( California-Nevada ) Atmospheric deposition of nitrogen and phosphorus in the annual nutrient load of Lake Tahoe ( California-Nevada ) contributes significant amo. water

(27)

26

Resour. Res. 30, 2207–2216. doi:10.1029/94WR00754.

Jeppesen, E., Brucet, S., and Eva, L. N. (2015). Ecological impacts of global warming and water abstraction on lakes and reservoirs due to changes in water level and related changes in salinity. Hydrobiologia 750, 201–227. doi:10.1007/s10750-014-2169-x.

Jeppesen, E., Søndergaard, M., Kronvang, B., Jensen, J. P., Svendsen, L. M., and Lauridsen, T. L. (1999). Lake and catchment management in Denmark. Hydrobiologia, 419–432. doi:10.1023/A:1017071602549.

Junger, P. C., Costa, F. C., Nobre, R. L. G., Kosten, S., Martins, E., Carvalho, F. De, et al. (2019). Effects of seasonality , trophic state and landscape properties on CO 2 saturation in low-latitude lakes and reservoirs. Sci. Total Environ. 664, 283–295.

doi:10.1016/j.scitotenv.2019.01.273.

Kitchell, J. F., Koonce, J. F., and Tennis, P. S. (1975). Phosphorus flux through fishes. SIL

Proceedings, 1922-2010 19, 2478–2484. doi:10.1080/03680770.1974.11896332.

Knoll, L. B., Hagenbuch, E. J., Stevens, M. H., Vanni, M. J., Renwick, W. H., Denlinger, J. C., et al. (2015). Predicting eutrophication status in reservoirs at large spatial scales using landscape and morphometric variables. Inl. Waters 5, 203–214. doi:10.5268/IW-5.3.812. Kristensen, P., Sondergaard, M., and Jeppesen, E. (1992). Resuspension in a shallow eutrophic

lake. Hydrobiologia 228, 101–109.

Mantzouki, E., Visser, P. M., Bormans, M., and Ibelings, B. W. (2015). Understanding the key ecological traits of cyanobacteria as a basis for their management and control in changing lakes. Aquat. Ecol. doi:10.1007/s10452-015-9526-3.

Matsuzaki, S. S., Usio, N., Takamura, N., and Washitani, I. (2007). Effects of common carp on nutrient dynamics and littoral community composition : roles of excretion and bioturbation. 168, 27–38. doi:10.1127/1863-9135/2007/0168-0027.

Menezes, R. F., Attayde, L., Kosten, S., Fernandes, R., Jose, M., Lacerot, G., et al. (2018). Differences in food webs and trophic states of Brazilian tropical humid and semi-arid shallow lakes : implications of climate change. Hydrobiologia.

Mitchell, N. L., and Lamberti, G. A. (2005). Responses in dissolved nutrients and epilithon abundance to spawning salmon in southeast Alaska streams. Limnol. Oceanogr. 50, 217– 227.

Naiman, R. J., Bilby, R. E., Schindler, D. E., and Helfield, J. M. (2002). Pacific Salmon , Nutrients , and the Dynamics of Freshwater and Riparian Ecosystems. Ecosystems 5, 399– 417. doi:10.1007/s10021-001-0083-3.

Nielsen, A., Trolle, D., Søndergaard, M., Lauridsen, T. L., Bjerring, R., Olesen, J. E., et al. (2012). Watershed land use effects on lake water quality in Denmark. Ecol. Appl. 22, 1187– 1200. doi:10.1890/11-1831.1.

Ometo, J. P. H. B., Martinelli, L. A., Victoria, M., Krusche, A. V, Victoria, R. L., and Williams, M. (2000). Effects of land use on water chemistry and macroinvertebrates in two streams of the Piracicaba river basin , south-east Brazil. Freshw. Biol. 44, 327–337.

(28)

27

Paerl, H. W. (2009). Controlling Eutrophication along the Freshwater – Marine Continuum : Dual Nutrient ( N and P ) Reductions are Essential. 593–601. doi:10.1007/s12237-009-9158-8.

Paul, M. J., and Meyer, J. L. (2008). “Streams in the urban landscape,” in Urban Ecology: An

International Perspective on the Interaction Between Humans and Nature (Springer), 207–

231. doi:10.1007/978-0-387-73412-5_12.

Polis, G. a, and Hurd, S. D. (1995). Extraordinarily high spider densities on islands: flow of energy from the marine to terrestrial food webs and the absence of predation. Proc. Natl.

Acad. Sci. U. S. A. 92, 4382–4386. Available at:

http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=41948&tool=pmcentrez&render type=abstract.

Ramankutty, N., Evan, A. T., Monfreda, C., and Foley, J. A. (2008). Farming the planet : Geographic distribution of global agricultural lands in the year 2000. Global Biogeochem.

Cycles 22, 1–19. doi:10.1029/2007GB002952.

Schaus, M. H., Vanni, M. J., Wissing, T. E., and Stein, R. A. (1997). Nitrogen and phosphorus excretion by detritivorous gizzard shad in a reservoir ecosystem. Limnol. Oceanogr. 42, 1386–1397.

Schindler, D. W. (2006). Recent advancement in understanding and management of eutrophication)_Schindler 2006. 51, 356–363.

Sereda, J. M., Hudson, J. J., Taylor, W. D., and Demers, E. (2008). Fish as sources and sinks of nutrients in lakes. Freshw. Biol. 53, 278–289. doi:10.1111/j.1365-2427.2007.01891.x. Seto, K. C., Güneralp, B., and Hutyra, L. R. (2012). Global forecasts of urban expansion to 2030

and direct impacts on biodiversity and carbon pools. PNAS 109, 16083–16088. doi:10.1073/pnas.1211658109.

Soballe, D. M., and Kimmel, B. (1987). A Large-Scale Comparison of Factors Influencing Phytoplankton Abundance in. Ecology 68, 1943–1954. doi:10.2307/1939885.

Sondergaard, M., Kristensen, P., and Jeppesen, E. (1992). Phosphorus release from resuspended sediment in the shallow and wind- exposed Lake Arreso , Denmark. Hydrobiologia 228, 91–99.

Subalusky, A. L., Dutton, C. L., Rosi, E. J., and Post, D. M. (2017). Annual mass drownings of the Serengeti wildebeest migration influence nutrient cycling and storage in the Mara River.

Proc. Natl. Acad. Sci. U. S. A. 114, 7647–7652. doi:10.1073/pnas.1614778114.

Subalusky, A. L., and Post, D. M. (2018). Context dependency of animal resource subsidies.

Biol. Rev., 1–22. doi:10.1111/brv.12465.

The World Bank (2019). Agricultural Land (% of land area). Food and Agriculture Organization, electronic files and web site. Available at:

https://data.worldbank.org/indicator/AG.LND.AGRI.ZS?type=points

Thornton, K. W. (1984). Regional comparisons of lakes and reservoirs : geology , climatology , and morphology. Lake Reserv. Manag. 1, 261–265. doi:10.1080/07438148409354521.

(29)

28

Vanni, M. J. (2002). Nutrient cycling by animals in freshwater ecosystems. Annu. Rev. Ecol.

Syst. 33, 341–370. doi:10.1146/annurev.ecolsys.33.010802.150519.

Vanni, M. J., Arend, K. K., Bremigan, M. T., Bunnell, D. B., Garvey, J. E., González, M. J., et al. (2005). Linking Landscapes and Food Webs : Effects of Omnivorous Fish and

Watersheds on Reservoir Ecosystems. Bioscience 55.

Vanni, M. J., Boros, G., and McIntyre, P. B. (2013). When are fish sources vs. sinks of nutrients in lake ecosystems? Ecology 94, 2195–2206.

Vanni, M. J., Bowling, A. M., Dickman, E. M., Hale, R. S., Higgins, K. a, Horgan, M. J., et al. (2006). Nutrient cycling by fish supports relatively more primary production as lake productivity increases. Ecology 87, 1696–709. Available at:

http://www.ncbi.nlm.nih.gov/pubmed/16922320.

Vanni, M. J., Renwick, W. H., Bowling, A. M., Horgan, M. J., and Christian, A. D. (2011). Nutrient stoichiometry of linked catchment-lake systems along a gradient of land use.

Freshw. Biol. 56, 791–811. doi:10.1111/j.1365-2427.2010.02436.x.

Vannote, R. L., Minshall, G. W., Cummins, K. W., Sedell, J. R., and Cushing, C. E. (1980). The River Continuum Concept. Can. J. Fish. Aquat. Sci. 37, 130–137. doi:10.1139/f80-017. Wang, J., Da, L., Song, K., and Li, B.-L. (2008). Temporal variations of surface water quality in

urban, suburban and rural areas during rapid urbanization in Shanghai, China. Environ.

Pollut. 152, 387–393. doi:10.1016/j.envpol.2007.06.050.

Williamson, T. J., Renwick, W. H., Vanni, M. J., Bremigan, M. T., and Conroy, J. D. (2018). The importance of nutrient supply by fish excretion and watershed streams to a eutrophic lake varies with temporal scale over 19 years. Biogeochemistry 6, 233–253.

(30)

29

Chapter 1 Precipitation, landscape properties and land use interactively affect

water quality of tropical freshwaters

Regina Lúcia Guimarães Nobrea, Adriano Calimana,b*, Camila Rodrigues Cabralc, Fernando de Carvalho Araújob_{, Joris Guérin}d_{, Fabíola da Costa Catombé Dantas}a_{, Letícia Barbosa Quesado}b_, Eduardo Martins Venticinquea,b_{, Rafael Guariento}e_{, André Megali Amado}f_{, Patrick Kelly}g_{, Mike} Vannih, Luciana Silva Carneiroa,b

aPrograma de Pós-Graduação em Ecologia, Universidade Federal do Rio Grande do Norte, Natal, RN, Brazil b Departamento de Ecologia, Universidade Federal do Rio Grande do Norte, Natal, RN 59078-900, Brazil c Departamento de Ciências do Mar, Universidade Federal de São Paulo, Santos, SP 11030-400, Brazil d Instituto de Computação, Universidade Federal Fluminense, Rio de Janeiro, RJ, Brazil

e Laboratório de Ecologia, Instituto de Biociências, Universidade Federal de Mato Grosso do Sul, MS, Brazil. f Departamento de Biologia, Universidade Federal de Juiz de Fora, Juiz de Fora, MG 36036-900, Brazil g Department of Biology, Rhodes College, Memphis, United States

h Department of Biology, Miami University, Oxford, Ohio, United States

*Corresponding author at: Departamento de Ecologia, Universidade Federal do Rio Grande do Norte, Natal, RN 59078-900, Brazil

Email Address: [email protected]

(31)

Precipitation, landscape properties and land use interactively affect

water quality of tropical freshwaters

Regina Lúcia Guimarães Nobrea, Adriano Calimana,b,⁎, Camila Rodrigues Cabralc,

Fernando de Carvalho Araújob, Joris Guérind, Fabíola da Costa Catombé Dantasa, Letícia Barbosa Quesadob, Eduardo Martins Venticinquea,b, Rafael Dettogni Guarientoe, André Megali Amadoa,f, Patrick Kellyg,

Michael J. Vannih, Luciana Silva Carneiroa,b a

Programa de Pós-Graduação em Ecologia, Universidade Federal do Rio Grande do Norte, Natal, RN, Brazil b

Departamento de Ecologia, Universidade Federal do Rio Grande do Norte, Natal, RN 59078-900, Brazil c

Departamento de Ciências do Mar, Universidade Federal de São Paulo, Santos, SP 11030-400, Brazil d_{Instituto de Computação, Universidade Federal Fluminense, Rio de Janeiro, RJ, Brazil}

e_{Laboratório de Ecologia, Instituto de Biociências, Universidade Federal de Mato Grosso do Sul, MS, Brazil.} f

Departamento de Biologia, Universidade Federal de Juiz de Fora, Juiz de Fora, MG 36036-900, Brazil g

Department of Biology, Rhodes College, Memphis, United States h

Department of Biology, Miami University, Oxford, OH, United States

H I G H L I G H T S

• Landscape features and precipitation can mediate land use effects on water quality.

• Human land use near shores of lakes and reservoirs decreases their water quality.

• Precipitation patterns mediate the ef-fects of land use on water quality. • Climate and geomorphology of lakes are

needed to understand land use effects. • Effective policies for climate change and

land use are vital to protect freshwater.

G R A P H I C A L A B S T R A C T

a b s t r a c t a r t i c l e i n f o

Article history:

Received 31 October 2019

Received in revised form 29 January 2020 Accepted 30 January 2020

Available online 31 January 2020 Editor: Ouyang Wei

Keywords: Shallow lakes Subsidy

Nonpoint source pollution Nitrogen

Phosphorus

Globally, conversion of pristine areas to anthropogenic landscapes is one of the main causes of ecosystem service losses. Land uses associated with urbanization and farming can be major sources of pollution to freshwaters pro-moting artiﬁcial inputs of several elements, leading to impaired water quality. However, how the effects of land use on freshwater quality are contingent on properties of the local landscape and climate is still poorly under-stood. The aim of this study was to evaluate the effects of landscape properties (morphometric measurements of lakes and their catchments), precipitation patterns, and land use properties (extent and proximity of the land use to freshwaters) on water quality of 98 natural lakes and reservoirs in northeast Brazil. Water quality im-pairment (WQI) was expressed as a composite variable incorporating parameters correlated with eutrophication including nitrogen (N), phosphorus (P) and Chlorophyll-a concentration. Regression tree analysis showed that WQI is mainly related to highly impacted“buffer areas”. However, the effects of land use in these adjacent lands were contingent on precipitation variability for 13% of waterbodies and on surface area of the buffer in re-lation to the volume of waterbody (BA:Vol) for 87% of waterbodies. Overall, effects on WQI originating from the

Science of the Total Environment 716 (2020) 137044

⁎ Corresponding author at: Departamento de Ecologia, Universidade Federal do Rio Grande do Norte, Natal, RN 59078-900, Brazil. E-mail address:[email protected](A. Caliman).

https://doi.org/10.1016/j.scitotenv.2020.137044

Contents lists available atScienceDirect

Science of the Total Environment

(32)

land use in the adjacent portion of the lake were ampliﬁed by high precipitation variability for ecosystems with highly impacted buffer areas and by high BA:Vol for ecosystems with less impacted buffer areas, indicating that ecosystems subjected to intense episodic rainfall events (e.g. storms) and higher buffer areas relative to aquatic ecosystem size (i.e. small waterbodies) are more susceptible to impacts of land use. Land use at the catchment scale was important for the largest ecosystems. Thus, ourﬁndings point toward the need for considering a holistic approach to managing water quality, which includes watershed management within the context of climate change.

1. Introduction

With the increasing human demand for natural resources caused by population and economic growth, Earth has experienced large scale transformations of its natural landscapes (Vitousek et al., 1997;Seto et al., 2012;Song et al., 2018;Marques et al., 2019). Indeed, human ac-tivities have modified a considerable proportion of natural ecosystems, for instance, farming lands, i.e. croplands and pastures (for livestock production), which occupy a combined area of approximately 34% of the Earth's ice free land surface (Ramankutty et al. 2008; Carpenter, Stanley and Vander Zanden, 2011). These changes have led to dramatic changes to the integrity of ecosystems, including their biodiversity and the maintenance of ecological functions and services they provide (Foley et al., 2005; IPBES, 2018). Freshwater systems are extremely valuable (Wilson and Carpenter, 1999), providing many ecosystem ser-vices such as human and animal water consumption, biodiversity main-tenance, biogeochemical cycles, carbon storage, hydrological regulation, irrigation,_{fishing and recreation (}Brasil et al., 2016;Carpenter et al., 1998;Daily et al., 1997;Schallenberg et al., 2013;Schindler, 2012). Al-though lakes, reservoirs, and rivers cover a combined area of only 2.3% of the inland surface, these ecosystems host at least 9.5% of the Earth's described animal species (Reid et al., 2019). However, they are among the most threatened ecosystems in the world (Dudgeon et al., 2006; Sala et al., 2000), experiencing higher rates of environmental degrada-tion and biodiversity loss compared to marine and terrestrial ecosys-tems (Reid et al., 2019). Most threats are directly or indirectly associated with anthropogenic-mediated land use changes in their catchments (Doubek et al., 2015;Jeppesen et al., 1999;Tromboni et al., 2019;Wang et al., 2008). Indeed, the downhill position of inland aquatic systems relative to their catchments facilitates theflux of alloch-thonous materials toward them by gravity (Vanni et al., 2005). There-fore, anthropogenic activities occurring in catchments are major sources of nonpoint pollution to freshwaters including enhanced inputs of several chemicals, such as nitrogen (N) and phosphorus (P) (Bennett et al., 2001;Paul and Meyer, 2008;Vanni et al., 2011).

Land use-mediated nutrient inputs are brought to water bodies via surface runoff, erosion and/or leaching (Foley et al., 2005). Anthropo-genic land use effects may promote cultural eutrophication and algal growth, thereby reducing water quality (Carpenter et al., 1998) and af-fecting ecosystem-level processes (e.g. lake productivity and nutrient cycling) and biodiversity (e.g. species richness, abundance and compo-sition) (Cabral et al., 2019;Doubek et al., 2015;Mantzouki et al., 2015;

Schindler, 2006). It can also modify the physical conditions of the aquatic environment, such as water temperature (LeBlanc et al., 1997) inorganic turbidity and water transparency (Declerck et al., 2006). In contrast, catchments with large proportions of land covered by forested areas and natural vegetation generally export lower amounts of dis-solved and particulate allochthonous materials to water bodies. This buffer effect is achieved through mechanisms such as plant nutrient up-take, reduced surface runoff and reduced soil erosion (Foley et al., 2005;

Lowrance et al., 1997;Mayer et al., 2007).

Several factors related to catchment land use can affect water qual-ity, including the type of land use (i.e. croplands, pasture, urbanization) (Tang et al., 2005;Vanni et al., 2001), its extent (i.e. the area of the catchment impacted) and its proximity to the aquatic system

(Declerck et al., 2006;Nielsen et al., 2012;Soininen and Luoto, 2012). For example, intensive agriculture can pollute aquatic ecosystems by the transport ofﬁne sediments, fertilizers, and pesticides (Feld et al., 2016a;Knoll et al., 2003). Extensive livestock production on pastures promotes soil compaction, changes in vegetation cover and produces manure, with consequences for soil biogeochemical cycles, as well as the links between soil processes and the concentrations of dissolved and particulate materials in aquatic systems (Neill et al., 2001). Finally, urbanization is especially detrimental to water quality where wastewa-ter treatment is deﬁcient (Foley et al., 2005) and impervious surfaces are common (Hobbie et al., 2017). However, it is not clear how water quality indicators may interactively depend on other factors such as the proximity of land use to the water body (Declerck et al., 2006;

Nielsen et al., 2012;Soininen and Luoto, 2012), geomorphological and topographic characteristics of catchments, morphological aspects of the aquatic system itself, and climate (Alahuhta et al., 2011;Hayes et al., 2015;Mattsson et al., 2005;Price, 2011). For instance, even if most of the catchment is anthropogenically impacted, it is possible that more pristine land cover in the portions adjacent to the lake could still buffer the impact of nutrients and sediment transport to the water body (Carpenter et al., 1998;Dodson et al., 2005;Muscutt et al., 1992). On the other hand, water quality can be severely affected by land use impacts concentrated in regions adjacent to the water body (Tran et al., 2010).

Landscape properties such as soil type, topography (e.g. slope), the absolute and relative size of catchments and their aquatic systems (Bremigan et al., 2008;Hayes et al., 2015;Vannote et al., 1980) and the origin (i.e. natural or artiﬁcial) of the aquatic systems (Dodson et al., 2006;Nielsen et al., 2012) are factors that affect the magnitude of transport rates of allochthonous material to waterbodies, as well as determine the sensitivity of these systems to external inputs of mate-rials. Therefore, landscape properties can also mediate, through com-plex ways, the effects of land use properties on aquatic systems. Finally, the magnitude and variability of precipitation can either inten-sify (e.g. via soil percolation or surface runoff) or attenuate (e.g. via di-lution) land use effects on aquatic systems (Jeppesen et al., 2015).

Combined long-term monitoring data for water quality and catchment land use dynamics are scarce (but see, e.g.Renwick et al., 2008, 2018;Richards et al., 2009), but an alternative ap-proach to understanding the complexity of interactions among the aforementioned mechanisms is to use a space-for-time ap-proach comprising aquatic systems that are distributed over large spatial scales (Soranno et al., 2017, 2019). The main goal of this study is to evaluate whether and how landscape properties and precipitation mediate the effects of land use properties on tropical inland aquatic systems in northeast Brazil. For this, we sampled 98 aquatic ecosystems distributed over 29,000 km2_encompassing

dif-ferent origins (i.e. natural lakes and human-made reservoirs), cli-mate (i.e. humid and semi-arid), landscape properties (i.e. the absolute and relative morphometric measurements of lakes, reser-voirs and their catchments), and land use patterns (i.e. extent and proximity of land cover to the water body) (Cabral et al., 2019;

Junger et al., 2019). This study has the potential to provide essen-tial information for decision making regarding land management and freshwater conservation.

(33)

2. Methods 2.1. Study area

This study was performed during the dry season of 2012 (Septem-ber) across 98 lentic water bodies located in the state of Rio Grande do Norte, northeastern Brazil (Fig. 1). These environments encompass perennial natural lakes (n = 30) and man-made artiﬁcial reservoirs (n = 68) formed by dams constructed between 1915 and 1950 by the National Department of Works for Drought Control (DNOCS, 2015). Their location ranges from the coastal to the semi-arid region, encompassing 14 (out of 16) watersheds of the state. Sixty-eight per-cent of the studied ecosystems are located in a semi-arid region (annual precipitation≈ 400–800 mm) with the remaining ecosystems distrib-uted along the coastal region within a sub-humid and humid climate (annual precipitation_{≈ 800–1200 mm) (BSh climate, Köppen} classiﬁ-cation). Most systems are shallow (90% with depthb 4 m) and have small surface area (89%b 1 km2_{) (}_{Table 1}_{). The majority of ecosystems}

in the semi-arid region are arti_{ﬁcial reservoirs (85%), while reservoirs} and natural lakes encompass 35% and 65% of aquatic systems distrib-uted across the coastal humid climate, respectively (Fig. 1). The spatial distribution of studied aquatic systems spans a gradient of catchment land use and trophic state, ranging mostly from pristine forested areas to agricultural land cover and from oligotrophic to hyper-eutrophic con-ditions (Table 1) (Cabral et al., 2019;Junger et al., 2019).

2.2. Field sampling and determination of the water quality parameters Impaired water quality is commonly associated with cultural eutro-phication, caused by the excessive inputs of N and P (Carpenter et al., 1998). Phytoplanktonic chlorophyll-a (Chl-a) has also been considered a proxy for productivity and trophic state of waterbodies (Boyer et al., 2009). Therefore, we used concentrations of N, P and Chl-a as proxies for water quality. To obtain concentrations of N, P and Chl-a for each lake, water samples were collected from the subsurface portion (ap-proximately 0.3 m deep) of the water column with a Van Dorn sampler. Samplings were performed at six sites in littoral and limnetic habitats (3 per habitat) and integrated into a single sample per water body. Water samples were analyzed in the laboratory for Chl-a, total nitrogen (TN) and total phosphorus (TP) (Valderrama, 1981). Determination of TN (mg L−1) was done by a carbon analyzer (TOC-V) coupled with a nitro-gen analyzer (VPN module) and an autosampler (Shimadzu). Total phosphorus (mg L−1) was determined by persulfate oxidation followed by measurement of reactive soluble phosphorus method (Murphy and Riley, 1962). Determination of Chl-a (μg L−1_{) was made according to}

(Jespersen and Christoffersen, 1987)

Because TP, TN, and Chl-a were all signiﬁcantly correlated with each other (Table 2), and for better interpretability, we created a“water qual-ity impairment” composite variable (hereafter, water quality impair-ment, WQI) by running a Principal Component Analysis (PCA) on the three response variables, (similarly toKnoll et al., 2015). We performed a broken-stick distribution to evaluate how many PCA axes were signiﬁcant.

2.3. Measurements of precipitation, landscape, and land use properties Lake morphometric characteristics, i.e. surface area and perimeter, were calculated through satellite image processing (delineation of poly-gon shapeﬁles) using ArcGIS 10.5 (ESRI, 2017). Lake depth was mea-sured at the central part of each lake using a calibrated rope, and lake volume was calculated through the hyperbolic function 0.43 × lake area × depth (Post et al., 2000). To analyze the importance of spatial scale (distance between the lake and the respective land use) on the ef-fects of land use on water quality, total catchment areas and buffer areas, were delineated for each lake. It is relevant to note that buffer areas in the context of this study depicts the near-freshwater land

zone around the water body comprising a distance of 100-meter from its shore, differently from the more classic use of the term“buffer zones_{” which deﬁnes a portion of permanently conserved natural} vege-tation along a watercourse (Karr and Schlosser, 1978;Muscutt et al., 1992). Delineation of catchments were based on lakes polygon shape_{ﬁles and it was performed using ArcHydro 2.0 Toolbox (}Arc Hydro Tools, 2011) on ArcGIS 10.5 (ESRI, 2017). Details on methodology for catchment area delineation can be found on supplementary material.

To avoid mismatches between the catchment area and buffer area, buffers were manually edited tofit within the catchment area of their specific lake. Morphometric landscape properties (catchment and buffer area) were calculated in ArcMap 10.5 (ESRI, 2017). Morphomet-ric characteristics as the catchment area-to-lake volume ratio (CA:Vol), buffer area-to-lake volume ratio (BA:Vol) and lake perimeter-to-lake volume ratio (LP:Vol) were also calculated as they are expected to in flu-ence the connectivity on the land-water interface (Table 3).

Slope (an indicator of steepness) of catchment and buffer area was also measured for each lake. We extracted slopes from a 30-m digital el-evation model (DEM) comprising the state of Rio Grande do Norte using the slope function at ARCMAP software. Data was gathered from the Brazilian Geomorphometric database (TOPODATA), available athttp:// www.dsr.inpe.br/topodata/acesso.php. However, because the region studied is mostlyﬂat, average slope of all catchments and buffers did not present a lot of variation (mean and SD; 2.1° ± 1.24°; 1.97° ± 0.77°, respectively) so we decided to not include this variable in further analysis.

Land use proportions relative to the year of 2012 were quantiﬁed for each catchment area and buffer area using data available through Pro-ject MapBiomas - Collection 3 of Brazilian Land Cover & Use Map Series (Projeto MapBiomas, 2018). Total area of each land use type was calcu-lated on ArcGIS 10.5. Subsequently, categories of land use provided by MapBiomas were grouped within two main groups: forested areas and anthropogenic areas. Anthropogenic use includes urban/developed zones and agriculture areas (cropland and pasture). They were grouped as these types of land use are expected to have similar effects on the in-crease of nutrient inputs on waterbodies (Foley et al., 2005). As forest land cover and anthropogenic land cover were strongly negatively cor-related, we used only anthropogenic land use in the model.

To address the temporal variability in precipitation present in the dataset, the magnitude and temporal variability of precipitation (here-after, Totprecipand Cvprecip, respectively) for each lake were computed.

There are seven weather stations distributed across the state of RN and for each one of them, the magnitude of precipitation was calculated as total monthly cumulative precipitation from September 2010 to Sep-tember 2012 (Source: INMET– Instituto Nacional de Meteorologia). To estimate precipitation temporal variability, the coefﬁcient of variation was obtained from monthly cumulative precipitation data for the same period. To obtain the magnitude and the temporal variability of precipitation for each lake, the weighted average of the values for the different stations was computed, with weights proportional to the dis-tance between the lake and the stations. Values obtained with the inter-polation approach are consistent with spatial variability in precipitation expected by the Worldclim long-term (1970–2010) (Fick and Hijmans, 2017; Fig. S2). More details about how the precipitation was computed can be found on the supplementary material.

2.4. Statistical analysis

We used a regression tree analysis (RTA) to verify how water quality impairment is related to precipitation, landscape, and land use proper-ties. RTA is a non-parametric statistical technique that allows one to ex-plain the variation of a single response variable by one or more explanatory variables (Breiman et al., 1984; DE’ath and Fabricius, 2000). Advantage of RTA over other multivariate tests is that it can un-ravel interactions among predictors and produces a straightforward

3 R.L.G. Nobre et al. / Science of the Total Environment 716 (2020) 137044

(34)

Fig. 1. Panel a show the studied lakes sampled on Rio Grande do Norte, Brazil. Natural lakes are represented by white dots and reservoirs by black dots. Land cover types were grouped according to the purposes of this study. Natural vegetated areas refers to natual forest, savanna and grassland formations. Anthropogenic land use encompasses farming (agriculture/pas-ture) and urban land uses. Original data was available through Project MapBiomas - Collection 3 of Brazilian Land Cover & Use Map Series. Map was build using ArcGIS 10.5.1 (ESRI, 2017). Panel b is presenting the average monthly precipitation from 2010 to 2012 in the studied lakes calculated from interpolation of the seven INMET automatic weather stations (stars). The dashed line depicts the semi-arid delimitation according to SUDENE, 2017. To the right of the line is the humid region and to the left is the semi-arid.